"Grid-Tape": A High-Throughput Platform for Brain Connectomics and Nanoscale Structural Analysis

脑回路的线路图是现代科学的基础和基本问题之一， 神经科学以突触分辨率提取电路大小的体积需要大规模、高通量 脑组织的许多薄切片的成像。研究人员的直接目标是捕获并绘制立方体 在突触分辨率下，大脑皮层的厚度为2.5毫米，需要收集和成像25，000个组织切片， 单一数据集。组织切片和成像的缓慢速度仍然是本领域的主要限制。 连接组学，需要许多年来获得这种大小的数据集。一些研究直觉（例如， 哈佛大学、珍妮莉亚研究院、艾伦研究所）正在积极寻求解决方案， 吞吐量成像。扫描电子显微镜（SEM）提供快速、可靠的样品切片和收集 通过自动胶带收集超薄切片机（ATUM-SEM），以及自动真空样品切割 通过串行块面（SBF-SEM）或聚焦离子束（FIB-SEM）。然而，SEM具有固有的差 分辨率和信号电平，这是由于扫描电子束像素收集的串行性质。在 相比之下，基于相机的透射电子显微镜（TEM）可以快速，同时收集 在比SEM更高的分辨率下，它可以达到数百万像素，但受到单个样品支架处理缓慢的限制。 为了克服这种吞吐量限制，哈佛大学的研究人员正在开发一种磁带样本 用于TEM的基板，称为“GridTape”，包括与以下兼容的带形式的薄膜覆盖的槽： 市售的ATUM设备。切片通过柱内卷对卷TEM成像进行成像 阶段哈佛选择Luxel作为他们的合作伙伴进行初步调查，并展示了~ 70倍的速度 用ATUM切割小鼠和果蝇脑薄切片拾取的改进。的 拟议的项目将优化和扩大GridTape，以满足大规模的迫切需求 连接组学研究。技术挑战包括卷到卷薄膜层压、膜破裂， 噪声和图像背景噪声。本建议的具体目标是：（1）开展试点 用于生产具有所需数量的成像槽的GridTape原型的制造方法 由研究人员（5，000 - 16，000）连续长度，生产吞吐量为2，000个插槽/天;（2） 优化GridTape规格以满足最终用户需求，将损坏插槽率降低至<1%;以及（3） Connectomics现场评估，在GridTape原型上采集数据集，具有> 10，000个插槽和 GridTape和SEM数据集之间的定量比较，证明了拾取速度和成像 生产能力至少与现有做法相当。第二阶段将通过更大的插槽解决生产规模扩大问题 计数和更低的成本每插槽。

The wiring diagram of brain circuits is one of the foundational and fundamental questions of modern neuroscience. Extracting a circuit-sized volume at synaptic resolution requires large-scale, high-throughput imaging of many thin sections of brain tissue. The immediate goal of researchers is to capture and map a cubic millimeter of cortex at synaptic resolution, requiring the collection and imaging of 25,000 tissue sections for a single dataset. The slow speed of tissue sectioning and imaging remains a major limitation to the field of connectomics, requiring many years to acquire a dataset of this size. A number of research intuitions (e.g. Harvard University, Janelia Research Campus, Allen Institute) are actively seeking solutions for higher- throughput imaging. Scanning electron microscopy (SEM) offers fast, reliable sample sectioning and collection via Automatic Tape-collecting Ultra-Microtome (ATUM-SEM), as well as automated in-vacuum sample cutting via Serial-Block-Face (SBF-SEM) or Focused Ion Beam (FIB-SEM). However, SEM has intrinsically poor resolution and signal level, due to the serial nature of the scanning electron beam pixel collection. In comparison, camera-based transmission electron microscopy (TEM) allows fast, simultaneous collection of millions of pixels at higher resolution than SEM, but is limited by slow handling of individual sample supports. To overcome this throughput limitation, researchers at Harvard University are developing a tape sample substrate for TEM called “GridTape”, comprising thin-film-covered slots in a tape form that is compatible with commercially available ATUM equipment. Sections are imaged via an in-column reel-to-reel TEM imaging stage. Harvard has chosen Luxel as their partner for the initial investigation and demonstrated ~70X speed improvement in pickup of thin sections of mouse and Drosophila melanogaster brains cut with an ATUM. The proposed project will optimize and scale up GridTape to meet the urgent unmet needs of large scale connectomics research. The technical challenges include reel-to-reel thin film lamination, film breakage, wrinkling, and image background noise. The specific aims of this proposal are: (1) Develop a pilot manufacturing process for the production of GridTape prototypes having the number of imaging slots required by researchers (5,000-16,000) in a continuous length, with production throughput of 2,000 slots/day; (2) Optimize GridTape specifications to meet end-user requirements, reducing damaged slot rate to <1%; and (3) Connectomics field evaluation, acquiring datasets on GridTape prototypes have >10,000 slots and a quantitative comparison between GridTape and SEM datasets demonstrating pickup speed and imaging throughput at least equal to existing practice. Phase II will address manufacturing scale-up with greater slot counts and lower cost per slot.